Patchwork connman: tist of connman do not support powerpc64

login
register
mail settings
Submitter chunrong guo
Date April 26, 2013, 8:07 a.m.
Message ID <1366963623-13626-1-git-send-email-B40290@freescale.com>
Download mbox | patch
Permalink /patch/48925/
State New
Headers show

Comments

chunrong guo - April 26, 2013, 8:07 a.m.
connman build for powerpc 64bit boards were producing errors like this:
    | make[1]: *** [plugins/plugins_tist_la-tist.lo] Error 1
    | make[1]: *** Waiting for unfinished jobs....
    | make: *** [all] Error 2
    | ERROR: oe_runmake failed

Signed-off-by: Chunrong Guo <B40290@freescale.com>
---
 meta/recipes-connectivity/connman/connman.inc |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)
ml@communistcode.co.uk - April 26, 2013, 9:33 a.m.
On 26/04/13 09:07, Chunrong Guo wrote:
>      connman build for powerpc 64bit boards were producing errors like this:
>      | make[1]: *** [plugins/plugins_tist_la-tist.lo] Error 1
>      | make[1]: *** Waiting for unfinished jobs....
>      | make: *** [all] Error 2
>      | ERROR: oe_runmake failed
>
> Signed-off-by: Chunrong Guo <B40290@freescale.com>
> ---
>   meta/recipes-connectivity/connman/connman.inc |    3 ++-
>   1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/meta/recipes-connectivity/connman/connman.inc b/meta/recipes-connectivity/connman/connman.inc
> index 589ece8..5389f3a 100644
> --- a/meta/recipes-connectivity/connman/connman.inc
> +++ b/meta/recipes-connectivity/connman/connman.inc
> @@ -20,10 +20,11 @@ DEPENDS  = "dbus glib-2.0 ppp iptables gnutls \
>               ${@base_contains('DISTRO_FEATURES', '3g','ofono', '', d)} \
>               "
>   
> -INC_PR = "r19"
> +INC_PR = "r20"
>   
>   TIST = "--enable-tist"
>   TIST_powerpc = ""
> +TIST_powerpc64 = ""
>   
>   EXTRA_OECONF += "\
>       ac_cv_path_WPASUPPLICANT=${sbindir}/wpa_supplicant \

Is there a reason that we need to specifically enable TIST at all? This 
is quoted from one of the developers on the ConnMan list:

> it is special hardware where we need this plugin for. As I said, I was considering to remove the plugin actually. So it is safe to > disable this plugin.

> By default this plugin is actually disabled. And we do not even build test it for the releases.

I think if anyone needs it, shout up; otherwise we should go with the default and disable it.

Cheers,
Jack.
Ross Burton - April 26, 2013, 9:36 a.m.
On 26 April 2013 10:33, Jack Mitchell <ml@communistcode.co.uk> wrote:
> I think if anyone needs it, shout up; otherwise we should go with the
> default and disable it.

Agreed - we can PACKAGECONFIG it and the machines that need it can enable it.

Ross
Koen Kooi - April 26, 2013, 9:46 a.m.
Op 26 apr. 2013, om 11:36 heeft "Burton, Ross" <ross.burton@intel.com> het volgende geschreven:

> On 26 April 2013 10:33, Jack Mitchell <ml@communistcode.co.uk> wrote:
>> I think if anyone needs it, shout up; otherwise we should go with the
>> default and disable it.
> 
> Agreed - we can PACKAGECONFIG it and the machines that need it can enable it.

Making connman machine specific which means PRSERV will go insane.
Ross Burton - April 26, 2013, 10:06 a.m.
On 26 April 2013 10:46, Koen Kooi <koen@dominion.thruhere.net> wrote:
> Making connman machine specific which means PRSERV will go insane.

Why would this cause prserv to go insane?  The only things build-time
depending on connman should be images so I don't see how it will cause
crazy rebuilds, so I must be missing something.

Ross
Martin Jansa - April 26, 2013, 10:19 a.m.
On Fri, Apr 26, 2013 at 11:06:07AM +0100, Burton, Ross wrote:
> On 26 April 2013 10:46, Koen Kooi <koen@dominion.thruhere.net> wrote:
> > Making connman machine specific which means PRSERV will go insane.
> 
> Why would this cause prserv to go insane?  The only things build-time
> depending on connman should be images so I don't see how it will cause
> crazy rebuilds, so I must be missing something.

I think it was meant as making it machine specific without changing it
to MACHINE_ARCH.

And I would rather build unused (on my MACHINEs) tist, then build it for
each MACHINE.

Using PACKAGECONFIG and enabling it by DISTRO_FEATURES or based on arch
is fine.
Koen Kooi - April 26, 2013, 10:21 a.m.
Op 26 apr. 2013, om 11:33 heeft Jack Mitchell <ml@communistcode.co.uk> het volgende geschreven:

> On 26/04/13 09:07, Chunrong Guo wrote:
>>     connman build for powerpc 64bit boards were producing errors like this:
>>     | make[1]: *** [plugins/plugins_tist_la-tist.lo] Error 1
>>     | make[1]: *** Waiting for unfinished jobs....
>>     | make: *** [all] Error 2
>>     | ERROR: oe_runmake failed
>> 
>> Signed-off-by: Chunrong Guo <B40290@freescale.com>
>> ---
>>  meta/recipes-connectivity/connman/connman.inc |    3 ++-
>>  1 files changed, 2 insertions(+), 1 deletions(-)
>> 
>> diff --git a/meta/recipes-connectivity/connman/connman.inc b/meta/recipes-connectivity/connman/connman.inc
>> index 589ece8..5389f3a 100644
>> --- a/meta/recipes-connectivity/connman/connman.inc
>> +++ b/meta/recipes-connectivity/connman/connman.inc
>> @@ -20,10 +20,11 @@ DEPENDS  = "dbus glib-2.0 ppp iptables gnutls \
>>              ${@base_contains('DISTRO_FEATURES', '3g','ofono', '', d)} \
>>              "
>>  -INC_PR = "r19"
>> +INC_PR = "r20"
>>    TIST = "--enable-tist"
>>  TIST_powerpc = ""
>> +TIST_powerpc64 = ""
>>    EXTRA_OECONF += "\
>>      ac_cv_path_WPASUPPLICANT=${sbindir}/wpa_supplicant \
> 
> Is there a reason that we need to specifically enable TIST at all? This is quoted from one of the developers on the ConnMan list:
> 
>> it is special hardware where we need this plugin for. As I said, I was considering to remove the plugin actually. So it is safe to > disable this plugin.
> 
>> By default this plugin is actually disabled. And we do not even build test it for the releases.
> 
> I think if anyone needs it, shout up; otherwise we should go with the default and disable it.

What's it actually doing? Kernel loads wifi firmware, you can use hciattach manually or uim automatically to enable bluetooth on wl1xxx chips, which are the only users of the TI shared transmission stuff. As far as I'm concerned it can be globally disabled :)

Patch

diff --git a/meta/recipes-connectivity/connman/connman.inc b/meta/recipes-connectivity/connman/connman.inc
index 589ece8..5389f3a 100644
--- a/meta/recipes-connectivity/connman/connman.inc
+++ b/meta/recipes-connectivity/connman/connman.inc
@@ -20,10 +20,11 @@  DEPENDS  = "dbus glib-2.0 ppp iptables gnutls \
             ${@base_contains('DISTRO_FEATURES', '3g','ofono', '', d)} \
             "
 
-INC_PR = "r19"
+INC_PR = "r20"
 
 TIST = "--enable-tist"
 TIST_powerpc = ""
+TIST_powerpc64 = ""
 
 EXTRA_OECONF += "\
     ac_cv_path_WPASUPPLICANT=${sbindir}/wpa_supplicant \